On the origin of the occasional spring nitrate peak in Greenland snow

Geng, L.; Cole-Dai, J.; Alexander, B.; Erbland, J.; Savarino, J.; Schauer, A. J.; Steig, E. J.; Lin, P.; Fu, Q.; Zatko, M. C.

doi:https://doi.org/10.5194/acp-14-13361-2014

Articles | Volume 14, issue 24

https://doi.org/10.5194/acp-14-13361-2014

© Author(s) 2014. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/acp-14-13361-2014

© Author(s) 2014. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 14, issue 24

Research article

|

16 Dec 2014

Research article |

| 16 Dec 2014

On the origin of the occasional spring nitrate peak in Greenland snow

L. Geng, J. Cole-Dai, B. Alexander, J. Erbland, J. Savarino, A. J. Schauer, E. J. Steig, P. Lin, Q. Fu, and M. C. Zatko

Download

Final revised paper (published on 16 Dec 2014)
Supplement to the final revised paper
Preprint (discussion started on 07 Apr 2014)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC C2059: 'Review Geng et al.', Anonymous Referee #1, 04 May 2014
- AC C3007: 'response to Reviewer 1', Lei Geng, 31 May 2014
RC C2285: 'Review of Geng et al.', Anonymous Referee #2, 10 May 2014
- AC C3008: 'response to Reviewer 2', Lei Geng, 31 May 2014

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Lei Geng on behalf of the Authors (20 Jun 2014) Author's response Manuscript

ED: Referee Nomination & Report Request started (03 Jul 2014) by Jan Kaiser

RR by Anonymous Referee #3 (14 Jul 2014)

Suggestions for revision or reasons for rejection

This manuscript advances an interesting, but very speculative, explanation for springtime peaks of nitrate in relatively recent Greenland snow (specifically Summit in this case). It has long been known that nitrate exhibited a broad summer maximum in preindustrial Greenland snow (as observed in firn and ice cores) but since ~ 1950 (perhaps as early as 1900 by some accounts) it has become difficult to recognize this annual cycle due to additional peaks that can occur at varied times of the year (e.g., Finkel et al., 1986; Whitlow et al., 1992; Yang et al., 1995; Burkhart et al., 2006). These earlier studies implied, or explicitly stated, that the extra nitrate peaks probably reflected transport of polluted airmasses to the Greenland ice, with scavenging by snowfall events and perhaps dry deposition of nitric acid and particulate nitrate enhancing snow nitrate concentrations.

In the present study, the depth profile of isotopic composition of nitrate in a shallow snowpit is presented (along with profiles of nitrate and selected other ion concentrations). The record covers approximately 3 full years, starting in May 2007 at the surface and apparently ending in summer 2004 at 2.1 m depth. Authors identify 5 nitrate peaks, noting that all or part of 4 summer peaks would be expected. They assert that the “extra” nitrate peak was deposited in spring of 2005 and then devote much of the manuscript to making a case that the enhanced concentrations of nitrate along with small decreases in delta O-18, cap delta O-17, and small increase in delta N-15 can all be linked to low ozone column coming out of the winter of 2005 causing increased local production, and deposition, of nitrate. The basic idea is that the smaller ozone column allows for higher UV flux, which could cause more photolysis of nitrate (also hydrogen peroxide) at depth in the snowpack, leading to enrichment of delta N-15 (note, this should also be at depth) in the snow and particularly a higher ratio of OH/O3 above the snow. Relative enhancement of OH could also partly result from greater photolysis of boundary layer O3. If more of the locally produced nitrate came from oxidation of NOx by OH (rather than O3) the oxygen isotopes would be expected to decrease as observed, if locally produced nitrate is a significant or dominant fraction of the total nitrate deposited at Summit.

Having laid out the basic argument based on the single “extra nitrate peak” in this snowpit, in the last 3.5 pages of the discussion the authors seek supporting evidence from an earlier 1-m deep (just over 1 year long) Summit pit and then extend their hypothesis by comparing other years with “spring” nitrate peaks in a firn core to the springtime O3 column in corresponding years.

One major problem that I have with the manuscript is the implicit assumption that much of the nitrate in snow at Summit, and apparently essentially all of the isotopic signature in that nitrate, results from locally produced nitrate. Hastings et al. (2004) argued that postdepostional processes had modest or no discernable impacts on the isotopic composition of nitrate preserved in firn (and subsequently ice) at Summit, a conclusion that was backed up by laboratory studies reported by McCabe et al. (2005 (JGR, 110, 2004JD005484)). This finding was further supported by extensive sampling of surface snow and shallow pits at Summit in summers of 2010 and 2011 when photochemistry is most active, (Fibiger et al. ,2013 (GRL, 40, grl.50659)) and no changes in the oxygen isotopes that could be explained by local processing were observed. As a result of this bias to explain nitrate concentrations and isotopic composition through local processes I feel that the authors were much to quick to dismiss the possibility that spikes outside the broad summer maximum might be related to long-range transport of pollution or possibly biomass burning smoke.

Note that it is possible that Hastings and her group are incorrect, and nitrate at Summit is dominated by local production. However, the discussion of the local processes that may be influencing nitrate in this manuscript is disjointed. The comments of reviewer 2 on the original manuscript made a lot of cogent points, which were dismissed (generally not in a convincing manner) in the authors’ response and largely ignored in this revised version. I agree with nearly all of the issues that referee raised and suggest they must be addressed before this paper should be published.

In particular, I agree with the comment that dating of both the snowpit and the firn core are critical to this story, and suspect in the present version. Starting with the pit, the entire hypothesis is based on assigning the nitrate peak at 1.5 m to spring of 2005 (peak 3’ in Fig. 1). Several weak lines of evidence suggest that this peak could just as plausibly be summer 2005, making the peak labeled 3 sometime later in summer, or even fall, of 2005. (Note that peak 3 has marked decreases in the oxygen isotopic ratios and small increase in delta N-15, and does not look all that much like summer peaks 1, 2 and 4 (especially according to the UW lab).) My case for suggesting that 3’ could be summer 2005 rests on the facts that 1) the depth separation between the “summer” nitrate peaks and “late-winter spring” Na peaks in 2007 and 2006 are quite small, much like the separation between 3’ and the 2005 Na peak, while nitrate peak 3 is about 45 cm higher in the pit, and 2) the depth interval between the partial summer peak (2004) at 2.1 m depth and peak 3 is at least 85 cm compared to 60 and 65 cm between the summer picks for 2005 to 2006 and 2006 to 2007, respectively. Granted, assigning 3’ to be summer 2005 makes the year between summers 2005 and 2006 the fat one, but less anomalously so since compaction increases with depth (annual layer thicknesses should be smaller near 2 m than in the top meter, if accumulation is roughly constant). Further support for a fat annual layer between summers 2006 and 2005 (my suggested redating) may be provided by the measurements of bamboo stakes presented in the supplemental material and in the response to reviewer 2, though it has to be noted that these are measurements of surface height change, not accumulation of snow, and include effects of compaction (See Dibb and Fahnestock, 2004 (JGR, 109, 2003JD004300)). Assuming ~constant compaction over each year, these measurements suggest the surface climbed (relative to the base of the poles) by 65 cm between Aug 2006 and Aug 2007, 80 cm 8/06 to 8/05 and 75 cm 8/05 to 8/04. Note, these complications interpreting the stake measurements make dating method B in the manuscript poorly constrained.

While it is not possible for me to say whether 2005 has nitrate peaks in spring and summer versus summer and fall with information I have, the authors ought to (have to) do a better job validating their subannual dating. If I am correct and the layer at 1.5 m is summer snow, the rest of the manuscript needs to be entirely reworked (or discarded). In a 2 meter snowpit, the visible stratigraphy clearly shows the difference between winter (finer grained) and summer snow (prominent hoar layers), likewise the density profile (lower in the summer). I fully agree with reviewer 2 that nitrate should not be used to identify summer, but excess Cl (or the Cl/Na ratio) is a very unambiguous summer marker (Whitlow et al., 1992; Dibb et al., 2007 (Atmos. Env., 41, j.atmosenv.2006.12.010). It is unfortunate that the Ca data from pit are no good since this is a great spring marker, though as the authors note the annual Na maximum is often just below or in the same layers as the April dust peak. I would be surprised if neither of the isotope labs involved in this study measured the oxygen isotopes in the snow (water) but the manuscript does not mention whether these data exist, these will also show useful winter and summer markers.

The dating of the firn core is even more troubling. While not very clearly described in the manuscript, the response to reviewer 2 makes it clear that “spring” nitrate peaks were identified solely by counting the frequently coincident Na and Ca peaks as annual layer markers and then counting nitrate peaks over the same depth interval. When the latter was larger, the difference was assigned to “spring” deposition regardless of where it may have fallen between a pair of Ca/Na annual markers. As pointed out by referee 2, three of the five example “springtime” nitrate peaks between 5 and 10 m are clearly below (before) both the Na and Ca peaks used to identify late winter/spring (Fig 2). Seems these would have to be early winter, or possibly fall events, yet they are assigned to the spring??? Authors state that it is not possible to resolve seasonality (or differential timing of peaks) in a core sampled at 3 cm resolution, yet one can clearly see this is not correct in the 2 5-m sections presented in this figure (even as deep as 45-50 m the summer nitrate peaks almost always clearly show up above (after) the spring Ca peaks). Similarly, earlier ice core studies mentioned above all seemed confident that they could tell summer nitrate spikes from those coming in at different times through the year in the younger part of the respective records (e.g., Finkel et al., 1986; Whitlow et al., 1992; Yang et al., 1995; Burkhart et al., 2006). As noted in relation to dating the pit, Ca and Cl/Na should provide quite dependable spring and summer marks against which to assess the timing of all nitrate peaks. Visible stratigraphy (light and dark bands) persist to depths equivalent to 50,000 year old ice at Summit, but may require a trained observer to identify with confidence (Alley et al., 1997 (JGR, 102, 96JC03837). If available, the water isotopes in the snow also help with subannual dating. Without much higher confidence that the authors have correctly identified spring nitrate peaks rather than just years with an extra peak due to plume transport, Figure 3 and all discussion of it has very little value.

Hide

ED: Reconsider after major revisions (15 Jul 2014) by Jan Kaiser

AR by Lei Geng on behalf of the Authors (11 Aug 2014) Author's response Manuscript

ED: Referee Nomination & Report Request started (15 Aug 2014) by Jan Kaiser

RR by Anonymous Referee #3 (29 Aug 2014)

Suggestions for revision or reasons for rejection

Geng et al. revision 2. The authors claim that this draft of the manuscript addresses the two main issues they feel I raised in regard to the previous version; dismissing long range transport of pollution as viable source of extra (outside of summer) nitrate peaks in the snow and firn records too quickly, and concern about dating of the pit and core. In my earlier review I also noted that I agreed with most of the comments raised by reviewer 2 on the initial submission, and that I felt these needed to be addressed in the manuscript (not just in the response to reviews).

This draft does directly address dating of the snowpit, and establishes the timing of the extra nitrate as a late winter/spring event. In fact, the authors claim that the peak is in snow they think accumulated in Feb (which is problematic for their hypothesis, more below). However, the authors choose not to attempt to date the firn core records, and continue to assume that any year with 2 nitrate peaks must have one in the spring and the other in the summer. They justify this decision on the basis of previous work by Cole-Dai et al who estimated that subannual dating of Summit firn cores was imprecise by up to +/- 4 months. I fully agree that it is not wise to assert that a peak in Na or Cl/Na can be dated to a specific week, or even month, but I am pretty certain that the winter peak in Na shows up in snow that accumulated in DJFM while the summer peak in Cl/Na occurs JJA. A “springtime” nitrate peak might show up in MAMJ in any given year, but it has to be between the winter and summer snow layers. As noted by reviewer 2, and myself, several of the extra nitrate peaks shown in Fig 2 (just 2 short sections of the firn core) fail this test and thus can not be springtime events. As a result, the comparison between O3 column variations and years with so-called springtime nitrate peaks (Fig 3 and discussion thereof) is not informative.

Authors dismiss the possibility that the nitrate peak in Feb/Mar 2005 resulted from long-range transport primarily due to the phase of the NAO, suggesting that meridional transport into the Arctic requires strong positive NAO index. It may be true that on average northward transport of mid latitude pollution is enhanced when the NAO is in positive phase, but advection of pollution plumes to Summit is a common event during winter and early spring. Kramer et al. (2004) identified 21 such events during DJFM in both 2008/09 and 2009/10 (42 winter/early spring transport events in just two years) using Flexpart and measurements of NOx, NOy, PAN and NMHC in the air just above the snow. Similarly, Duderstadt et al., (2014) show numerous peaks of nitrate in the surface snow at Summit during this same season in 2000/01 and 2001/02, and indicate that the NCAR WACCM model does a reasonable job of simulating these via tropospheric transport of pollution from the mid latitudes. It should be noted that the NAO index was negative in 3 out of these 4 winters. (As an aside, I also point out that Kramer et al. (2014) clearly show that PAN is the dominant component of the NOy that reaches Summit, accounting for a low of 50% in mid summer and up to 80% in April.)

Regarding the hypothesis that the spring 2005 nitrate is locally produced (recycled) from snow sourced NOx and OH, the Feb/Mar timing of this peak is kind of problematic. There is not a whole lot of sunlight at Summit at that time of year, for example the midday measured J NO3- on 25 Mar 2004 was just 0.6 x 10^-8, and both NO and OH were close to detection limits in the air 1 m above the snow. In early Feb the sun is still below the horizon for more than 20 hours each day (sunrise approx. 27 Jan). Therefore it is hard to imagine that O3 depletion could enhance the UV flux enough to stimulate significant release of NOx and OH for much of Feb and Mar. Of course, the solar elevation does increase quickly so far north (by 13 Apr 04 the midday J NO3- was almost 4 times higher than on 25 Mar) but relaxation of the polar vortex and replenishment of the O3 column is usually pretty well advanced by April. The TUV calculations presented for PAN and nitrate photolysis should have been a hint that photolysis of nitrate (and also hydrogen peroxide as additional source of OH) in snow at Summit in Feb would be expected to be quite weak (1.3 times a very small number is still very small). Even if I am mistaken and there are enough photons hitting the snow in March to recycle nitrate, why should it stop and preserve a spring peak? The actinic flux continues to increase to much higher levels through April, May and much of June, so why would locally formed nitrate deposited in Feb/Mar not continue to photolyze and get redeposited at the new surface, eventually merging with the usual summer peak? Fibiger et al. (2013) suggest that very little of the initially deposited nitrate ever gets photolyzed, such that the recycled fraction is also a very minor component of preserved nitrate.

As noted, reviewer 2 raises additional concerns with this hypothesis that I still feel have not been adequately addressed.

Duderstadt, K. A., J. E. Dibb, C. H. Jackman, C. E. Randall, S. C. Solomon, M. J. Mills, N. A. Schwadron, and H. E. Spence (2014), Nitrate deposition to surface snow at Summit, Greenland following the 9 November 2000 solar proton event, Journal of Geophysical Research 119, 6938-6957, doi:10.1002/2013JD021389.

Kramer, L. J., D. Helmig, J. F. Burkhart, A. Stohl, S. Oltmans, and R. E Honrath (2014), Seasonal variability of atmospheric nitrogen oxides and non-methane hydrocarbons at the GEOSummit station, Greenland, Atmospheric Chemistry and Physics Discussion, 14, 13817-13867, www.atmos-chem-phys-discuss.net/14/13817/2014/doi:10.5194/acpd-14-13817-2014.

Hide

ED: Reconsider after major revisions (03 Sep 2014) by Jan Kaiser

AR by Lei Geng on behalf of the Authors (13 Nov 2014) Author's response Manuscript

ED: Publish subject to technical corrections (13 Nov 2014) by Jan Kaiser

AR by Lei Geng on behalf of the Authors (13 Nov 2014) Author's response Manuscript

Download

Article (288 KB)
Metadata XML

Short summary

Examinations on snowpit and firn core results from Summit, Greenland suggest that there are two mechanisms leading to the observed double nitrate peaks in some years in the industrial era: 1) long-rang transport of nitrate and 2) enhanced local photochemical production of nitrate. Both of these mechanisms are related to pollution transport, as the additional nitrate from either direct transport or enhanced local photochemistry requires enhanced nitrogen sources from anthropogenic emissions.